Every Major Google Algorithm Update (1998 to 2026): A Working Timeline

Published 27 May 2026 · 13 min read

Most Google algorithm timelines on the internet were written by US publications watching US SERPs. Australian SERPs moved on different timing for several major updates, and Australian small businesses felt different things. This is the comprehensive timeline of every confirmed major Google algorithm update from PageRank to AI Mode, annotated with what The SEO Company actually saw on .com.au sites since opening in 2007. By the end you will have a single reference that names the update, dates it, explains what it did, and gives the honest Australian read on whether it mattered for SMB SEO at the time.

How to read this timeline

The "major" filter is doing some work here. Google confirms several thousand ranking changes per year. Most are invisible. A handful per year shift rankings hard enough that an agency notices a wobble across multiple clients in the same week. Those are the ones in this timeline.

Each entry has the same shape: name, confirmation date, what it changed, and the Australian impact where we can speak to it. Pre-2007 entries are reconstructed from the public record because The SEO Company did not exist yet. From 2007 onward, the notes are from work the agency actually ran. We say so where the observation is anecdotal versus where it shows up cleanly in client data.

The timeline groups by era because the updates clustered that way. Pre-Panda was the wild era. 2011 to 2013 was the penalty era. 2013 to 2018 was the semantic and mobile era. 2018 to 2022 was the quality era. 2023 onward is the AI era, and the cadence has accelerated.

The pre-Panda era (1998 to 2010)

PageRank (1998) was the original. Larry Page's link-counting paper became the engine. PageRank was the entire conversation for the first six years of SEO as a discipline. Reciprocal links, link wheels, paid directory packages, and PageRank "sculpting" with nofollow were the agency lingua franca of the era.

Florida (November 2003) was the first update that visibly punished aggressive SEO. Florida hit keyword stuffing and over-optimised anchor text. The Australian SMB market in 2003 was still mostly Yellow Pages-led for local lead generation, so the SEO casualties were the early adopters who had over-optimised. Agencies that survived Florida learned to write for humans first. Agencies that did not survive Florida went back to print.

Big Daddy (2005 to 2006) rebuilt the index infrastructure. Less a "ranking" update, more a foundation. Big Daddy hardened the URL canonicalisation logic and started to deflate the value of low-quality directory backlinks. Anyone running a serious SEO business in 2006 felt the shift in how directory links suddenly counted for less.

Universal Search (May 2007) was when Google started blending news, images, video, local, and books into the main results. For Australian local businesses, this was the moment that "appearing in Google" no longer meant the ten blue links. It meant the map pack, the image strip, and the local listings. The agency was three months old when this rolled out, and our first conversation with most clients shifted from "rank in the top ten" to "rank in the top three boxes that matter".

Vince (February 2009) was a brand-favouring update that quietly boosted big sites for big terms. Vince is often misunderstood. It did not penalise small business. It re-weighted ambiguous queries toward authoritative sites, which usually meant brands. Australian SMBs saw this most clearly on category-style queries like "running shoes Perth", where retailer brands started outranking small specialist shops.

Caffeine (June 2010) rebuilt the indexing system to enable near-real-time inclusion. The visible effect was speed. The structural effect was that fresh content could rank within minutes rather than weeks. The agency adjusted to this by changing the publishing cadence on a couple of news-style retailer clients, and we saw the speed of indexing improve materially that month.

The penalty era (2011 to 2013)

Panda (24 February 2011) changed everything. Panda was the first update with a clear "thin content" mandate. Sites with scraped, syndicated, or shallow articles dropped overnight. Panda was rolled into the core algorithm in early 2016, but its DNA still drives quality scoring.

The Australian impact of Panda was lighter than the global trade-press coverage suggested at the time. Most of the worst-affected sites globally were US content farms, which were less of a force in the local market. Where we saw Australian sites hit hard, it was usually the dirt-cheap local-directory operators with thousands of templated suburb pages and very thin content per page. A few never recovered.

Page Layout Algorithm (January 2012), sometimes called "Top Heavy", targeted pages with too many ads above the fold. Limited Australian footprint because Australian SMB sites usually had no above-the-fold ads to start with.

Penguin (24 April 2012) was the link-quality penalty. Penguin punished sites with manipulative backlink profiles. We had a handful of pre-Panda client sites that had paid for directory link packages in 2009 and 2010. Most of them survived Penguin because the Australian directory ecosystem of that era was less spammy than the global one. The sites that took the worst hits had bought into international link networks. Recovery work after Penguin was the first time an Australian agency could charge real money for "disavow file management".

Exact Match Domain Update (28 September 2012) stripped the ranking boost from low-quality EMDs. The "perthplumber.com.au" effect, where you bought an exact-match domain and instantly ranked, ended that week. We saw two client sites lose 40 to 60 per cent of organic traffic on EMDs that had previously coasted. Both rebuilt over the following year by earning links to the domain rather than relying on the URL.

Knowledge Graph (May 2012) was not strictly an algorithm update but it changed the SERP layout. Right-hand panels appeared for entities and brands. Wikipedia, ASIC business records, and the AU equivalent of structured data started mattering for entity recognition.

The semantic era (2013 to 2018)

Hummingbird (26 September 2013) was the move toward intent-based ranking. Hummingbird could parse longer conversational queries and try to understand what the searcher meant rather than just matching keywords. The agency's content briefing changed that quarter. We stopped asking copywriters for "the keyword phrase three times in the first paragraph" and started asking for "all the related sub-questions a customer would ask about this product or service".

Pigeon (24 July 2014) was the local search update. Pigeon tied local rankings more closely to traditional web search signals. This was a watershed for Australian local SEO. Before Pigeon, the local pack could be dominated by sites with thin content but strong directory citations. After Pigeon, the same sites needed actual website authority too. Our service offer expanded that year from "GBP and citations" to a properly integrated local SEO programme.

HTTPS as a Ranking Signal (6 August 2014) was a light boost for SSL-enabled sites. Most Australian SMB sites moved to HTTPS over the following 18 months, with the laggards finally migrating around the time Chrome started showing "Not Secure" labels in 2018.

Mobilegeddon (21 April 2015) punished sites that were not mobile-friendly. This was the busiest week for emergency development work in the agency's history. Many Australian trade businesses, particularly older builders and electricians, had been running Joomla or older WordPress sites with no responsive theme. The week before mobilegeddon, our delivery team did nothing else but emergency mobile theme swaps and viewport meta fixes. Within six weeks most affected clients had recovered baseline mobile rankings. The lasting effect was that we started insisting on mobile-first design from the first conversation with every new client.

RankBrain (26 October 2015) introduced machine learning into the ranking system. The third most important ranking signal at announcement, per Google. RankBrain handled queries Google had never seen before. The agency-facing impact was that "exact keyword" optimisation continued to lose value. Topic depth and intent matching mattered more.

Possum (1 September 2016) changed local ranking for businesses near the edge of a city limit. Some local Perth businesses with addresses in Cottesloe or Subiaco saw their rankings into the Perth CBD radius improve overnight. Others saw their rankings flatten because they were now being filtered alongside competitors with similar addresses. We rewrote the citation strategy for half a dozen local-service clients that quarter.

Intrusive Interstitials Update (10 January 2017) punished pop-ups and large interstitial ads on mobile. Mostly hit news sites and big publishers. Limited Australian SMB impact, though a handful of clients with newsletter sign-up pop-ups did need to redesign them as bottom-anchored banners.

Fred (8 March 2017) was unofficially named and unofficially confirmed. Fred targeted low-quality ad-heavy content sites. The agency lost no clients to Fred. We had no clients running that model.

The quality era (2018 to 2022)

Medic Update (1 August 2018), the first widely-named broad core update, hit health, wellness, finance, and other YMYL (Your Money or Your Life) topics. Quality scoring tightened around E-A-T signals (Expertise, Authoritativeness, Trustworthiness). Two of our healthcare practice clients took a 30 per cent hit on organic traffic in the four weeks after Medic. Recovery work meant building out genuine author bios, professional registration links to AHPRA, and explicit citation of peer-reviewed sources in clinical content. The traffic came back over six months.

Mobile-First Indexing (rolled out 2018 to 2020) flipped the default index from desktop to mobile. By March 2021, all sites were mobile-first indexed. Australian SMB sites mostly handled this fine because the mobilegeddon work in 2015 had forced the shift early.

BERT (October 2019) was the natural language processing model rolled into the ranking system. BERT improved Google's handling of prepositions and conversational queries. For long-tail SEO this was huge. The agency saw a measurable increase in long-tail capture across content-led clients over the following quarter.

Featured Snippet Deduplication (22 January 2020) stopped sites from ranking in both the featured snippet and the regular ten blue links. Featured snippets became more valuable but cost a top-ten ranking. Strategic question on every featured snippet target after that: is this snippet worth losing the underneath ranking?

Core Web Vitals as Ranking Signal (15 June to 31 August 2021) rolled CWV (LCP, FID, CLS) into the Page Experience signal. The signal weight was small but the publicity was huge. Half our 2021 client work that year was Core Web Vitals remediation, particularly on legacy WordPress and Magento builds. Australian e-commerce sites lagged the US benchmark on mobile LCP for most of 2021. We saw a clear gap in our agency-wide PageSpeed data, with median mobile LCP on AU client sites still over 4 seconds in mid-2021 when the global median was closer to 3.5.

MUM (May 2021) was the multimodal AI system announced at I/O 2021. MUM did not roll out as a single update. It started informing features over time and laid the groundwork for AI Overviews.

Product Reviews Update (April 2021, then iterated) raised the bar for product review content. Affiliate sites with thin reviews dropped. AU comparison sites took collateral damage. Sites with first-hand product testing notes, original photography, and explicit pros and cons survived and grew.

Spam Updates (multiple, late 2021 onward) consistently swept low-quality content and link spam. Routine for sites with clean profiles.

Helpful Content Update (25 August 2022) was the explicit "content for humans, not search engines" update. The HCU classifier became a site-wide signal in March 2024 (more below). The 2022 launch hit AU affiliate review sites the hardest. Local service businesses with first-hand content barely noticed.

SpamBrain (October 2022) improved spam detection and worked alongside the link spam updates. The era of "buying 200 forum profile backlinks for fifty dollars on Fiverr" ended for good around here.

The AI era (2023 to 2026)

March 2023 Core Update was the first big core update after ChatGPT's launch. It hit AU sites that had started using GPT-3.5-generated content at scale. We saw two affected clients in our book of work and recommended a full content audit, deletion of AI-only pages, and a rebuild of the underlying topical clusters with human-written briefs.

Hidden Gems Update (2023, soft launch) was Google's attempt to surface forum content, blog comments, and small-publisher commentary. Reddit, Quora, and niche forums climbed. The agency adjusted by including forum participation in selected B2B client strategies for the first time in five years.

AI Overviews / Search Generative Experience (May 2024) rolled AI-generated summaries into the SERP. Information queries dropped traffic to source pages by 30 to 40 per cent on the worst-affected sites in our book of work over the eight weeks after rollout. Commercial queries were less affected because AI Overviews mostly refuse to recommend a specific business. The structural lesson: traffic from definitional searches ("what is X") is now mostly Google's. Traffic from transactional searches is still yours to compete for.

March 2024 Core + Spam Update was the largest core update in Google's history. The Helpful Content classifier became a site-wide demotion signal. Spam updates removed sites engaged in scaled content abuse, site reputation abuse, and expired domain abuse. The cleanup ran across May and June 2024. We had no clients hit directly because none of them were running AI content at scale, but several of our clients' direct competitors lost ground that quarter.

Site Reputation Abuse Policy (5 May 2024) punished publisher sites that hosted third-party content irrelevant to the main site's purpose, often coupon and affiliate inserts. Some Australian news publishers were affected. Limited SMB impact.

November 2024 Core Update was a return to standard cadence. Modest movement across our book.

AI Mode (May 2025 onward) moved beyond AI Overviews to a full conversational interface inside Search. Question, follow-up, refinement, all without leaving Google. The AI Overview hit on information traffic deepened. Brand authority, entity recognition, and direct mentions in trusted publications became the new ranking levers. Pure SEO became pure AI-search visibility work.

March 2025 Core and August 2025 Core Updates continued to refine the quality signals introduced through 2024. By late 2025 the agency was saying out loud that "ranking" was no longer the right unit of work. The right unit was "being cited", whether by ten blue links or by AI Overviews or by AI Mode answers.

What the Australian footprint actually looks like

Pulling back from the timeline, three patterns recur in the AU client data across all 19 years of updates.

First, AU SMB sites lag the US benchmark on technical readiness by about 12 to 18 months. Mobilegeddon, HTTPS, CWV, mobile-first indexing each landed harder here because more local sites were still on legacy stacks. The remediation work was always available revenue for an agency willing to do it cleanly.

Second, the AU local SEO market has historically been less spammy than the US equivalent, which means penalty updates (Penguin, EMD, HCU, March 2024) did less damage on average to genuine AU SMB sites. The big losers were almost always either international content farms or AU sites that had bought into international tactics.

Third, every era's update set retired a class of low-effort SEO tactic and rewarded a higher-effort one. Florida killed keyword stuffing. Panda killed thin content. Penguin killed paid links. EMD killed lazy domain purchases. Mobilegeddon killed desktop-only design. HCU killed scaled mediocre content. AI Overviews are killing "what is X" traffic and rewarding direct first-hand expertise. The pattern continues.

What we believe

The right way to read 19 years of Google updates is not as a list of disasters to avoid. It is as a slow filter that progressively removes shortcuts and rewards original work. Every update in this timeline has been Google saying the same thing in a different register: do the actual work, write for humans, link because the link is useful, and run a site you would still run if Google did not exist. The agencies and clients that did that in 2007 are mostly still running. The ones that did not, are mostly not. The next ten years of updates will not change that pattern.

If you want a clean read on how your own site holds up against current core update signals, grab a free SEO audit. We will tell you what we see, including the parts you did not want to hear.

Related reading

• What is SEO? The pillar overview
• AI Search: how visibility works now
• Digital marketing for Australian businesses: a 2026 field guide
• Core Web Vitals and conversion

FAQ

What was the first major Google algorithm update?

Florida, in November 2003, was the first algorithm update that visibly penalised aggressive SEO. Florida targeted keyword stuffing and over-optimised anchor text. Before Florida, ranking was largely a function of PageRank and on-page repetition.

How many major Google algorithm updates have there been since 1998?

More than 30 confirmed major updates that materially changed rankings. Google ships several thousand smaller ranking changes per year, but only a handful per year are big enough that an agency notices a wobble across multiple clients in the same week. This timeline covers the 30-plus updates with that level of impact.

Did Australian websites get hit harder by certain Google updates?

Australian SMB sites lagged the US benchmark on technical readiness by roughly 12 to 18 months, which meant Mobilegeddon in April 2015, Core Web Vitals from 2021, and the mobile-first indexing rollout each landed harder in the Australian market. The remediation work was always available revenue for agencies willing to do it cleanly.

What was the biggest Google algorithm update since 2020?

The March 2024 Core and Spam Update was the largest core update in Google's history. The Helpful Content classifier became a site-wide demotion signal, and several spam updates removed sites engaged in scaled content abuse, site reputation abuse, and expired domain abuse. The cleanup ran across May and June 2024.

How do you recover from a Google algorithm update penalty?

It depends on the update. After Penguin in April 2012, recovery meant disavowing manipulative backlinks. After Medic in August 2018, it meant building genuine author bios and citing peer-reviewed sources. After the March 2024 HCU integration, it meant removing thin AI-generated content and rebuilding topical clusters with human-written briefs. The pattern is consistent: identify what the update was targeting, then remove the targeted behaviour from the site.